In logic, a clause is an expression formed from a finite collection of literals (variables or their negations) that is true either whenever at least one of the literals that form it is true (a disjunctive clause, the most common use of the term), or when all of the literals that form it are true (a conjunctive clause, a less common use of the term). That is, it is a finite disjunction or conjunction of literals, depending on the context. Clauses are usually written as follows, where the symbols are literals:
A clause can be empty (defined from an empty set of literals). The empty clause is denoted by various symbols such as , , or . The truth evaluation of an empty clause is always . This is justified by considering that is the neutral element of the monoid .
Every nonempty clause is logically equivalent to an implication of a head from a body, where the head is an arbitrary literal of the clause and the body is the conjunction of the negations of the other literals. That is, if a truth assignment causes a clause to be true, and none of the literals of the body satisfy the clause, then the head must also be true.
This equivalence is commonly used in logic programming, where clauses are usually written as an implication in this form. More generally, the head may be a disjunction of literals. If are the literals in the body of a clause and are those of its head, the clause is usually written as follows:
- If n = 1 and m = 0, the clause is called a (Prolog) fact.
- If n = 1 and m > 0, the clause is called a (Prolog) rule.
- If n = 0 and m > 0, the clause is called a (Prolog) query.
- If n > 1, the clause is no longer Horn.